Dilution susceptibility testing of 139 Shigella isolates from patients with acute shigellosis showed that all were inhibited by concentrations of 2.5μg of furazolidone per milliliter, with the majority being inhibited by 0.625μg/ml or less. Nineteen infants and children hospitalized for severe shigellosis were treated with 8 mg/kg/day of furazolidone for five days. Persistence of positive stool cultures for more than 48 hours or relapse after initial bacteriologic response was observed in 13 patients (68%). Clinical failure occurred in eight patients (42%). There was no decrease in the numbers of Shigella per gram of feces shed during therapy compared with pretreatment quantitations. The clinical and bacteriologic results of furazolidone therapy are comparable to the natural course of illness seen in placebo-treated patients and markedly inferior to those obtained with ampicillin therapy. On the basis of these findings and a review of previous reports we conclude that furazolidone is ineffective and should not be recommended for treating shigellosis.